AD is a neurodegenerative disorder that is characterized by memory loss and cognitive decline. AD is increasingly viewed as a disorder of cerebral glucose metabolism, which is known to start up to two decades before the clinical manifestation of AD. Several consensus statements have hence highlighted the need for metabolic biomarkers to differentiate AD induced brain changes from other forms of dementia, especially since pharmaceutical agents to treat AD are emerging. Highly specific and sensitive quantitative biomarkers of abnormal glucose metabolism, which can facilitate longitudinal studies, are urgently needed for the early detection and management of Alzheimer?s disease (AD). The main goal of this proposal is to develop a practical and reproducible J-resolved magnetic resonance spectroscopic imaging (MRSI) framework on a 7T scanner to probe abnormal metabolism in early AD subjects. The focus of the parent grant is to develop a short protocol for multi-parametric (T1, LGE, CINE) cardiac MRI using non-Cartesian sequences & machine learning. In this supplement, we propose to extend the computational framework and non- Cartesian sequences introduced in the parent grant to exploit the higher SNR offered by ultra- high field magnets (e.g 7T) to overcome the challenges in MR spectroscopic imaging. This approach will enable the accurate estimation of the concentration of J-coupled metabolites that have significant overlap in 1-D MRSI. Our previous supplement is facilitating the specificity and spatial resolution of T1rho, which is a metabolic biomarker for acidosis resulting from accumulation of waste products in AD. This proposal, if successful, will enable us to complement the acidosis biomarker with a broader array of metabolites which will enable us to characterize the various stages of the pathological cascade in early AD. Thus, this supplement will facilitate the development of a metabolic MRI toolbox, which will greatly strengthen research on AD and related dementias.